Circadian Characteristics of Sleep Propensity Function In Healthy Elderly: A Comparison With Young Adults

Transcription

1 Sleep 20(4): American Sleep Disorders Association and Sleep Research Society Sleep In Normal Subjects Circadian Characteristics of Sleep Propensity Function In Healthy Elderly: A Comparison With Young Adults Iris Haimov and Peretz Lavie Sleep Laboratory Faculty of Medicine Technion-Israel Institute of Technology Haifa Israel Summary: Changes in sleep-wake patterns are among the hallmarks of biological aging. Elderly persons complain of daytime drowsiness and difficulties in initiating and maintaining sleep. The question of whether age-related changes in sleep-wake distribution are the result of a dimunition in amplitude of the endogenous circadian pacemaker resulting in a decline in nocturnal sleep tendency and an increase in diurnal sleep tendency or a manifestation of the impact of the medical and psychosocial burden on sleep has not yet been fully determined. In the present study we utilized a 7/13 ultrashort sleep-wake paradigm to investigate the 24-hour sleep propensity function (SPF) in healthy elderly persons. Seventeen healthy elderly males aged years and eight young males aged participated in the study. All elderly subjects were living independently in the community and were vigorous physically active and socially engaged. The young adult subjects were students living on campus all with the same daily schedule. As anticipated polysomnographic measures from the night prior to experimental periods differed between the elderly and young subjects. Specifically the elderly had a reduction in percentage of sleep stage 3/4 and in sleep efficiency. The results of the 7/13 ultrashort sleep-wake paradigm showed that although aging did not affect the overall structure of the SPF there was an age-related trend toward lower circadian amplitude and advanced phase. Our findings suggest that these age-related changes in sleep propensity may contribute to the difficulties in initiating sleep and to the early morning awakening. Key Words: Healthy elderly-sleep-wake rhythms-napping-7/13 paradigm-normal young adults. Biological aging is marked by changing sleep-wake patterns: complaints of difficulties in initiating and maintaining sleep and of daytime drowsiness are more common in the elderly than in any other age group. These disturbances are recognized as being among the most common health problems of the elderly (1). As a result of sleep patterns that accompany normal aging healthy elderly require more time to fall asleep awaken more frequently during the night and wake up earlier in the morning than young adults. They have little or no slow-wave sleep and have decreased total sleep time (2-6). Coupled with these nighttime changes an increase in undesirable daytime naps which may further disrupt the individual's daily routine has been reported (7-10). The sleep of healthy elderly individuals appears to be redistributed over the 24-hour day; they spend less time in nocturnal sleep and relatively more time in daytime napping. This Accepted for publication February Address correspondence and reprint requests to Peretz Lavie Sleep Laboratory Faculty of Medicine Gutwirth Building Technion-Israel Institute of Technology Haifa Israel. 294 means that their total 24-hour sleep time is often equal to that of younger people. Because daytime sleepiness in late life has an enormous impact on quality of life level of functioning and ability to remain independent (2) preserving sleep integrity in late life is an important health priority. It is not yet clear however whether age-related changes in sleep-wake distribution are the result of a dimunition in the amplitude of the endogenous circadian pacemaker in which nocturnal sleep tendency declines and diurnal sleep tendency increases or a manifestation of the impact of the medical and psychosocial burden on sleep. In the present study we utilized an ultrashort sleepwake paradigm to test whether daytime sleepiness is an unavoidable consequence of the aging process. Subjects were instructed to attempt sleep for 7 minutes every 20 minutes during a 24-hour period. The total amount of sleep obtained in each trial was used to construct the 24-hour sleep propensity function (SPF) which describes the speed of falling asleep at different times across the 24-hour day. This schedule does not allow a significant accumulation of sleep which could obscure the temporal structure of sleep propensity. ~)

2 SPF IN HEALTHY ELDERLY COMPARED WITH YOUNG ADULTS Since its introduction the ultrashort sleejr-wake paradigm has been used to measure the SPF of young adult subjects (11-15). The temporal structure of the SPF of normal young adults is characterized by three distinct features: a midafternoon secondary peak in sleepiness an early evening nadir in sleepiness (the "forbidden zone" for sleep) and a nocturnal precipitous drop in sleep latency (the "nocturnal primary sleep gate"). The characteristics of the SPF were found to have a high intrasubject stability (15). The timing of the nocturnal sleep gate was shown to be a reliable circadian phase marker; it significantly correlated with the onset of noctornal secretion of melatonin (1617) responded by delay to evening exposure to bright light (18) and significantly differentiated between morning and evening persons (19). The specific purpose of the present study was to investigate the SPF of elderly healthy male subjects as determined by the ultrashort sleejr-wake paradigm and to compare it to the temporal structure of the SPF of young adults. Subjects METHOD AND DESIGN Seventeen healthy elderly males aged years and eight young males aged participated in the study. All elderly subjects were living independently in the community and were vigorous physically active and socially engaged. Young adult subjects were students living on campus all having the same daily schedule. All subjects were in good clinical condition and had no sleejr-wake complaints. None of the subjects suffered from sleep apnea dementia or depression or were taking any medication that could affect sleep or the noradrenergic system for at least 1 month prior to the study. None used alcohol on a regular basis and none was a habitual smoker. All subjects were individually interviewed and completed several questionnaires including a I-week sleep log a mini-sleep questionnaire (20) and the Technion sleep questionnaire to exclude sleep disturbances. For habituation purposes and to rule out sleep apnea syndromes and periodic limb movements all spent a night in the laboratory with polysomnographic recording approximately 1 week before the start of the study. To control subjects' sleep behavior prior to the study all subjects kept a I-week sleep log during the week preceding the study. On the basis of the sleep diaries young subjects slept 6.5 ± 1.5 hours per night with bedtimes between 2330 and 0130 hours. These are typical bedtimes and total sleep times for students and are similar to the data of young adults previously studied in our laboratory. Elderly subjects slept 6.0 ± 1.0 hours per night with bedtimes between 2230 and 0100 hours. Experimental design The ultrashort sleejr-wake paradigm has been previously described in detail (13); therefore only a brief description is provided here. The experimental period started after the subjects slept in the laboratory during the night. Subjects arrived at the sleep laboratory at 2100 hours before the start of the ultrashort sleejrwake paradigm. During the night sleep was recorded using standard measures of electroencephalography (EEG) electrooculography (EOG) and electromyography (EMG). All sleep records were scored in 30-second epochs according to standard criteria (21). Subjects slept in the laboratory from hours; at 0640 hours they were awakened. Electrodes were examined to ensure secure placements and at 0700 hours the subjects started an ultrashort sleejr-wake paradigm that lasted for 24 consecutive hours. During the experimental period the subjects were instructed every 20 minutes to lie in bed with eyes closed and to attempt to fall asleep. For each subject all sleep attempts were made in the same bedroom which was sound-attenuated air-conditioned and completely darkened. Electrophysiological recordings of electroencephalogram electro-oculogram and electromyogram were conducted during the 7-minute sleep attempts. At the end of the 7-minute period whether asleep or awake the subject was taken out of the bedroom for 13 minutes. In total the experimental period included 72 cycles of 7-minute sleep attempts followed by l3-minute intervening awake periods outside the bedroom. Polysomnographic recordings performed during each 7 -minute sleep attempt were conventionally scored for sleep stages. The total amount of sleep [stage 2 stage 3/4 and rapid eye movement (REM) sleep] obtained in each 7-minute trial represents sleep propensity at that particular time. During the 13-minute awake periods the subjects were free to converse read or write. Food and drinks (light snacks and noncaffeinated soft drinks) were available every 1-3 hours throughout the experimental period. Subjects were not isolated from time cues. To obtain a formal assessment of the time of the abrupt drop in nocturnal sleep latency (the nocturnal sleep gate) we used the method described in Lavie (l3). Previous results have shown that in most subjects tested with the ultrashort sleejr-wake paradigm nocturnal sleep latency decreased abruptly reaching within one-trial minimal levels of minutes that remained relatively constant until the end of the experimental period. The nocturnal pattern of sleep propensity following the gate was widely different from the

3 296 I. HAIMOV AND P. IA VIE TABLE 1. Nocturnal sleep data (means and standard deviations) for the nights preceding the experimental periods of young and old subjects Young Old Sleep measures (n = 8) (n = 17) P Sleep latency 23.7 (19.3) 22.6 (18.4) Total sleep time (minutes) (38.4) (33.4) % Sleep efficiency 85.8 (9.8) 75.4 (11.6) <0.04 REM latency (minutes) (38.6) 73.7 (46.9) % Sleep stage (10.2) 19.7 (9.8) % Sleep stage (1.2) 1.1 (0.63) % Sleep stage (6.7) 48.4 (10.2) % Sleep stage 3/ (4.5) 13.2 (7.7) <0.05 REM 19.1 (4.2) 17.6 (6.7) REM rapid eye movement; not significant. pattern of sleep propensity observed before the opening of the gate. It was characterized by high levels of sleepiness and low trial-to-trial variability. Therefore according to this method the critical trial (the nocturnal sleep gate) was defined as the first trial after 1900 hours in which total sleep exceeded 3.S minutes and which was followed by at least five out of six trials meeting the same criterion. Statistical analysis Nocturnal sleep measures for the two groups were compared by Mann-Whitney test (two-tailed) for total sleep sleep stages 1 2 3/4 and REM. To assess the main effects of age group on the overall level of sleepiness the average amount of sleep summed across the 72 trials and the average amount of sleep stages 1 2 3/4 and REM were compared by Mann-Whitney test (two-tailed). To assess differences in the 24-hour distribution of sleep propensity repeated measures analysis of variance (ANOV A) was performed on the means of total sleep and sleep stages 1 and 2 in 2-hour blocks. Since relatively few trials contained either REM or stage 3/4 these data were not included in the analysis. To assess the main effect of age on the circadian rhythm of sleep propensity a fifth-order polynomial curve was fitted to each subject's total sleep data (22). Each subject's sleep propensity minimum (corresponding to the forbidden zone for sleep) and maximum values (corresponding to the nocturnal sleepiness crest) and corresponding clock times were obtained from these curves and the slope between the minimum and maximum points was computed. Group means were compared by two sample t tests (twotailed). A formal assessment of the effects of age on the nocturnal sleep gates was also obtained by twosample t tests (two-tailed). Given the small sample size of young men and their substantial variability to comment on the possibility of a type 2 error when a between-group difference was TABLE 2. Total sleep time and stages 123/4 and rapid eye movement of the young and old subjects (means and standard deviations) in the 7/13 paradigm TST Stage 1 Stage 2 Stage 3/4 REM Young 2.18 :!: :!: :!: :!: :!: 0.1 Old 1.90 :!: :!: :!: 0.6 O.O:!: ::':: 0.02 z p <0.02 TST total sleep time; REM rapid eye movement; not significant. not detected and to assess the reliability of the results differences detectable at 80% power and O.OS significance were computed. This was done by using the means and variances calculated from these experiments for all variables in which parametric tests were used. Nocturnal sleep RESULTS Table 1 summarizes sleep data from nocturnal sleep periods for the nights preceding the experimental periods of young and old subjects. Mann-Whitney test revealed statistically significant differences between the young and elderly subjects in two measures: sleep efficiency and percentage of sleep stage 3/4 (8S.8 ± 9.8% vs. 7S.4 ± 11.6%; Z = 2.29 P < 0.04; 19.1 ± 4.S% vs ± 7.7%; Z = 2.10 P < O.OS respectively). Sleep propensity function Levels of sleep propensity Table 2 presents the average amounts (in minutes per trial) of sleep stages 1 2 3/4 and REM and of total sleep time (TST) of the young and elderly subjects. Mann-Whitney test revealed no significant effects of age on mean TST or on the means of stages 1 2 and 3/4. There was a statistically significant difference only in the mean of stage REM sleep time (Z = 2.29 P < 0.02). However this latter finding is clinically insignificant since relatively few trials contained REM sleep. Figure 1 depicts the mean sleep propensity function for the two age groups and Figs. 2a and b present the mean curves of sleep stages 1 and 2 calculated separately for the young and elderly subjects. As can be seen the SPFs of both age groups varied in a prominent circadian manner and the overall structure of sleepiness was similar. Compared with the elderly young subjects had more sleep at the beginning and.l

4 SPF IN HEALTHY ELDERLY COMPARED WITH YOUNG ADULTS 297 SLEEP PROPEITY FUNCTION STAGE 1 SLEEP 20~ ~ 'J ~J 6 GROUP YOUNG - ELDERLY 5 " I ** 23 HOUR OF DAY 3 "... ' " * * 7 FIG. 1. Mean SPFs of the two age groups. The curves were constructed by averaging total sleep time (sum of stages I 2 3/4 and REM sleep) in each of the 72 trials. Curves were smoothed by cubic spline. Asterisks indicate significant age differences for these blocks (post-hoc Bonferroni test). the end of the experiment and their evening forbidden zone for sleep was considerably more pronounced. Repeated ANOV A performed on the means of total sleep and sleep stages 1 and 2 in 2-hour blocks revealed no significant effect of age but a significant main effect of blocks [F(1l253) = p < ; F(lI253) = 9.52 p < ; F(lI253) = p < respectively; Greenhouse-Geisser Epsilon = respectively] and a significant interaction between age and blocks for total sleep and for sleep stage 1 [F(ll 253) = 3.82 p < ; F(II253) = 4.43 p < respectively]. Post-hoc Bonferroni tests showed that total sleep and sleep stage 1 during the first and last blocks of the experimental period (block hours and block hours) were significantly lower in elderly compared to young subjects. In contrast TST and sleep stage 1 during the early night period (block hours; the time of the forbidden zone for sleep) were significantly lower in the younger subjects. Power computations revealed that a difference of 1.1 and 0.5 minutes for total sleep and stage 1 respectively could be detected.!1l I":iI 15 ~ 10 Z... ~ 0.5 GROUP YOUNG - ELDERLY. \ \ \ 11 "' \ \ I * * 23 HOUR OF DAY STAGE 2 SLEEP a. * * 7 {O GROUP YOUNG - ELDERLY b. 3.0!1l 25 I":iI ~ 20 Z ~ '...=;..._ '---' ' HOUR OF DAY FIG. 2. Mean curves of sleep stages I and 2 for the young and old subjects. Curves were smoothed by cubic spline. Asterisks indicate significant age differences for these blocks (post-hoc Bonferrani test). TABLE 3. Minimum and maximum values and their respective clock times and the slopes of the lines connecting them derived from the fifth-order polynomial curve fitting \'! Circadian characteristics and polynomial analysis Two of the elderly subjects did not have an identifiable sleep propensity function minimum and were therefore eliminated from the polynomial analysis. Table 3 shows the mean clock hour and the minimum and maximum values of total sleep for each group as well as the mean slopes. Independent two-sample t tests (two-tailed) revealed a weaker rhythmic trend and Young Old (n = 8) (n = 15) P Maximum trial 5:20 (0:54) 4:24 (0:58) Maximum value (minutes) 5.79 (1.53) 4.39 (1.22) Minimum trial 19: 18 (2:06) 19:120:44) 0.12 Minimum value (minutes) 0.72 (0.86) 1.09 (0.98) Slope 0.17 (0.05) 0.12 (0.03) not significant.

5 298 I. HAIMOV AND P. LA VIE 7 ~ ~ Elderly ~ Young 6 5 "-I ~4 = ~ O~~rrrrrrrnTTTTTTTTTTlTrrfuTrl"'!TfulTTTITTTlrTTTlTTTTTTTTTTlTlTTTlm' Hour of Day FIG. 3. Fifth-order polynomial curves fitted to the SPFs of the two age groups. an earlier and lower SPF maximum in the elderly group compared with the younger subjects (Fig. 3). There was a significant difference between young and elderly subjects in the slope of the line connecting the maximum and minimum points (0.17 ± 0.05 and 0.12 ± 0.03 respectively; t = 3.22 P < 0.004). Likewise there were significant differences between the young and the elderly in the mean clock hour (0520 hours ± 54 minutes and 0424 hours ± 58 minutes respectively; t = 2.26 P < 0.04) and in maximum values of total sleep (5.79 ± 1.53 and 4.39 ± 1.22 respectively; t = 2.40 P < 0.03). There were no significant differences between the young and the elderly at the minimum point in the mean clock hour (1918 hours ± 126 minutes and 1912 hours ± 104 minutes respectively) or in the minimal level (0.72 ± 0.86 and 1.09 ± 0.98 respectively). Power computations revealed that a difference of 1 hour in the maximum and 2 hours 10 minutes in the minimum could be detected. Detectable differences in minutes of total sleep per 7 minutes were 1.6 and 1.0 minutes respectively. Despite the fact that the level of sleep propensity function of the elderly subjects compared to the young ones was lower in the number of blocks during the day the timing of the nocturnal sleep gate of the two groups was the same (0100 hours ± 180 minutes vs hours ± 140 minutes). Power computations revealed that a detectable difference of 3 hours could be found for the sleep gate. There were no correlations between nocturnal sleep stages and sleep stages during the experimental period (Spearman correlation) in either group. DISCUSSION In the present study we investigated the temporal structure of sleep propensity in elderly men. Specifi- cally the study sought to examine age effects on the circadian characteristics of the SPF without unwanted effects of illness. Therefore the subjects included in this study probably represent only the high-functioning healthy elderly male rather than the entire elderly population. Furthermore since there is now considerable evidence supporting gender differences in sleep and circadian rhythms of elderly subjects (2324) these findings cannot be generalized to apply to women. Polysomnographic measures from the night before the experimental periods predictably differed between this healthy elderly sample and the young subjects. Specifically elderly subjects had a reduction in the percentage of sleep stage 3/4 and in sleep efficiency. Nevertheless our data showed that aging did not affect the overall amount of day time sleep or the general shape of the SPF but only attenuated its circadian amplitude and advanced it. In the elderly as in the young subjects there was a minor midaftemoon peak centered at about 1500 hours (the secondary sleepiness peak) a nadir in sleepiness in the early evening (forbidden zone for sleep) and a decline in sleep latency during the night after the opening of the sleep gate. These changes however were more prominent in the young adults. The major differences between the groups were in the amplitude of the SPF: 1) a reduction in sleepiness in the elderly subjects compared to the young subjects during the morning hours ( hours) and at the end of the night ( hours). 2) an increase in sleepiness in the elderly subjects compared to the young ones during the early night period i.e. the time of the forbidden zone for sleep ( hours). These differences were also evident in the polynomial analysis. Elderly subjects have a less steep increase in nocturnal sleepiness lower peak sleep propensity levels and more advanced peaks. This latter finding may also explain the lower morning sleep propensity in the elderly. As can be seen in Fig. 1 sleep propensity in the young was still high at 0700 hours the last trial of the study whereas in the elderly sleep propensity declined at about 0500 hours. Interestingly although young subjects had more prominent forbidden zones for sleep there were no significant differences with respect to minimum values and their times. This may be explained by the elimination of two of the elderly subjects from the polynomial analysis because they did not have an identifiable SPF minimum. The elimination of these two subjects would tend to minimize group differences. Supporting data have been recently reported by Monk et al. (25) who found that subjective alertness rhythms of the elderly (especially elderly men) are less robust than those of the young showing lower amplitude and phase-advanced alertness rhythms... :

6 SPF IN HEALTHY ELDERLY COMPARED WITH YOUNG ADULTS 299 ~) 1/> We suggest that these age-related changes in sleep propensity may contribute to two central sleep complaints of the elderly. The reduction in the ability to maintain sleep during the early morning hours ( hours) may be the cause of the early morning awakenings and the increase in sleepiness during the early night period ( hours) which brings about a less dramatic opening of the sleeping gate may be the cause of the difficulties in initiating sleep. Nevertheless our data show that aging did not affect sleepiness in the midafternoon. This finding suggests that complaints of daytime drowsiness in the elderly cannot be the direct result of the age-related changes in sleep propensity but rather results from other factors such as changes in lifestyle. Likewise studies of daytime napping in old age support pathogenic roles for medical affective and cognitive disorders (26-29). The study of Buysse et al. (30) which examined temporal patterns of unintentional sleep episodes during 36 hours of constant wakeful bed rest in healthy elderly and young subjects is congruent with our findings. They found that using their study protocol reduced circadian variation in sleep tendency in elderly men compared with young adults. There have been relatively few prior polysomnographic studies of daytime sleep architecture in the healthy elderly. Using the multiple sleep latency test paradigm Richardson et al. (31) reported similar levels of daytime sleepiness and diurnal variability of sleepiness in college students and the elderly. Using the same paradigm Levine et al. (32) and Reynolds et al. (33) found diminished daytime sleepiness in healthy elderly compared with young adults. Contradictory data were reported by Carskadon (34) who found that healthy elderly subjects have shorter daytime sleep latencies compared to young adults. Our findings showing an age-related trend toward lower circadian amplitude and advanced phase support the notion of a primary age-related change in the mammalian circadian pacemaker. Several lines of evidence suggest that the circadian pacemaker oscillates more rapidly and with a lower amplitude in older than in younger animals and human beings. First the freerunning period of the rest-activity rhythm shortens with age in both animals (3536) and humans (37). Second animal studies have shown an age-related reduction in the amplitude of the rest-activity and sleepwake cycles (38). Our findings suggest that even though the elderly have less ability than young adults to sleep at night (including less ability to generate slow-wave sleep) they do not need to compensate during naps. Likewise our data show no correlation between nocturnal sleep stages and sleep stages during the experimental period in either group. In any case the observation of decreased slow-wave sleep and sleep efficiency in the nighttime sleep of the healthy elderly does not seem to imply that they are at greater risk for excessive daytime sleepiness. In this context it is possible that the habitual napping in the elderly represents volitional rather than obligatory napping. Acknowledgements: We are grateful to Arnir Zvuluni for his contribution to the control group data to Paula Rerer for her statistical assistance and to Gay Natanzon for her typing and editorial help. REFERENCES 1. Ford DE Kameron DB. Epidemiologic study of sleep disturbances and psychiatric disorders: an opportunity for prevention? lama 1989;262: Fowies OG. A profile of the Americans Washington D.C.: American Association of Retired Persons Dement WC Miles LE Carskadon MA. "White paper" on sleep and aging. lam Geriatr Soc 1982;30: Moran MG Thompson TL Nies AS. Sleep disorders in the elderly. Am 1 Psychiatry 1988:145; Prinz PN Vitiello MV Roskind MA Michael JT. Geriatrics: sleep disorders and aging. N Engl 1 Med 1990;23: Monk TH Reynold CF Buysse OJ et al. Circadian characteristics of healthy 80-year-olds and their relationship to objectively recorded sleep. 1 Gerontol 1991;46:M Buysse OJ Browman KE Monk TH Reynolds CF Fasiczka AL Kupfer OJ. Napping and 24-hour sleep/wake patterns in healthy elderly and young adults. lam Geriatr Soc 1992;40: Tune GS. Sleep and wakefulness in 509 normal human adults. Br 1 Med Psychol 969;42: Jacobs D Ancoli-Israel S Parker L Kripke DF. 24-h sleepwake patterns in a nursing home population. Psychol Aging 1989;4: Regestein QR Morris J. Daily sleep patterns observed among institutionalized elderly residents. 1 Am Geriatr Soc 1987;35: Lavie P Scherson A. Ultrashort sleep-wake schedule. I. Evidence for ultradian rhythmicity in sleepability. Electroencephalogr Clin NeurophysioI1981;52: Lavie P Zomer J. Ultrashort sleep-wake schedule. II. Relationship between ultradian rhythms in sleepability and the REM NONREM cycles and the effects of circadian phase. Electroencephalogr Clin Neurophysiol 1984;57: Lavie P. Ultrashort sleep-wake schedule. III. "Gates" and "forbidden zones" for sleep. Electroencephalogr Clin Neurophysiol 1986;63: Lavie P. To nap perchance to sleep--ultradian aspects of napping. In: Oinges D Broughton RJ eds. Napping: biological psychological and medical aspects. New York: Raven Press 1989: Lavie P. The 24-h sleep propensity function (SPF): practical and theoretical implications. In: Monk T ed. Sleep sleepiness and performance. Baffins Lane: Wiley 1991: Tzischinsky 0 Shlitner A Lavie P. The association between the nocturnal sleep gate and nocturnal onset of urinary 6-sulphatoxymelatonin. 1 Bioi Rhythms 1993;8: Shochat T Luboshitsky R Lavie P. Melatonin onset precedes the primary sleep gate. 1 Sleep Res 1996;Suppl 1:213 (abstract). 18. Tzischinsky 0 Lavie P. The effect of evening bright light on next-day sleep propensity. 1 Bioi Rhythms 1997 (in press). 19. Lavie P Segal S. Twenty-four-hour structure of sleepiness in morning and evening persons investigated by ultrashort sleepwake cycle. Sleep 1989;12: Zomer J Peled R Rubin A-H Lavie P. Mini-sleep questionnaire

7 HAIMOV AND P. LA VIE (MSQ) for screening large populations for EDS complaints. In: Koella Wp Ruther E Schulz H eds. Sleep '84. Stuttgart: Gustav Fischer Verlag 1985: Rechtschaffen A Kales A eds. A manual of standardized terminology techniques and scoring system for sleep stages of human subjects. Los Angeles: Brain Information ServicelBrain Research Institute University of California at Los Angeles Lack LC Lushington K. The rhythms of human sleep propensity and core body temperature. J Sleep Res 1996;5: Hoch CC Reynolds CF Monk TH et al. Comparison of sleepdisordered breathing among healthy elderly in the seventh eighth and ninth decades of life. Sleep 1990;13: Moe KE Prinz PN Vitiello MV Marks AL Larsen LH. Healthy elderly women and men have different entrained circadian temperature rhythms. JAm Geriatr Soc 1991;39: Monk TH Buysse DJ Reynolds III CF Kupfer DJ Houck PRo Subjective alertness rhythms in elderly people. J Biol Rhythms 1996;11: Gislason T Almqvist M. Somatic diseases and sleep complaints. Acta Med Scand 1987;221: Morgan K Healey DW Healey PJ. Factors influencing persistent SUbjective insomnia in old age: a follow-up study of good and poor sleepers aged Age Ageing 1989;18: Ford DE Kamerow DB. Epidemiologic study of sleep disturbances and psychiatric disorders. An opportunity for prevention? JAMA 1989;262: Bliwise DL. Sleep in normal aging and dementia. Sleep 1993; 16: Buysse DJ Monk TH Reynolds CF Mesiano D Houck PR Kupfer DJ. Patterns of sleep episodes in young and elderly adults during a 36-hour constant routine. Sleep 1993;16: Richardson GS Carskadon MA Orav GJ Dement We. Circadian variation of sleep tendency in elderly and young adult subjects. Sleep 1982;5:S Levine B Roehrs T Zorick F Roth T. Daytime sleepiness in young adults. Sleep 1988; 11 : Reynolds CF Jennings JR Hoch CC et al. Daytime sleepiness in the healthy "old-old": a comparison with the young adults. JAm Geriatr Soc 1991;39: Carskadon MA. Ontogeny and human sleepiness as measured by sleep latency. In: Dinges DF Broughton RJ eds. Sleep and alertness. New York: Raven Press 1989: Pittendrigh CS Daan S. Circadian oscillations in rodents: a systematic increase in their frequency with age. Science 1974;186: Morin LP. Age-related changes in hamster circadian period entrainment and rhythm splitting. J Biol Rhythms 1988;3: Weitzman ED Moline ML Czeisler CA Zimmerman Je. Chronobiology of aging: temperature sleep-wake rhythms and entrainment. Neurobiol Aging 1982;3: Welsh DK Richardson GS Dement WC. Effect of age on the circadian pattern of sleep and wakefulness in the mouse. J Gerontol 1986;41 : '1).).\